Psychosis represents a major global health concern. Emerging evidence implicates dysregulation of the hypothalamic-pituitary-adrenal (HPA) axis and altered cortisol profiles in the emergence and course of psychotic disorders. This protocol describes a planned systematic review that will synthesise observational evidence to address two primary objectives: (1) evaluate the association between cortisol levels and psychosis in individuals with confirmed psychotic disorders, and (2) examine cortisol levels in people identified as being at elevated or clinical risk for developing psychosis.
The review aims to consolidate empirical findings from case-control and cohort studies to clarify whether circulating or sampled cortisol measurements differ between patients and controls, and whether alterations are evident in prodromal or ultra-high-risk groups compared with lower-risk comparators. The protocol emphasises observational study designs that report original cortisol data, to map existing evidence without restricting by year or language of publication.
Electronic literature searches will be carried out in major bibliographic databases: Web of Science, PubMed, Scopus, Embase and PsycINFO. The authors will develop a primary search strategy in Web of Science and subsequently adapt that strategy to each of the other databases. The search uses a PECO framework (Population, Exposure, Comparator, Outcome) to structure terms and filters.
The initial PECO-derived search will employ Medical Subject Headings (MeSH) and equivalent database terms and phrases, exemplified by: ('Cortisol' OR 'Hydrocortisone') AND ('Psychosis' OR 'Psychotic disorders' OR 'Schizophrenia') AND ('HPA axis' OR 'Stress response') AND ('Clinical high risk' OR 'Prodromal' OR 'Ultra-high risk'). Additional phrases will be identified from relevant literature and incorporated as appropriate.
As a supplementary step, the review team will search Google Scholar and examine the first 1,000 publications returned to crosscheck database results and identify potentially missed studies. The combined approach is intended to be comprehensive for observational studies reporting cortisol measurements in the target populations.
Inclusion criteria
Exclusion criteria
If specific cortisol assay methods, sampling times, or matrices differ between studies, these details will be recorded during data extraction to inform interpretation and any subgrouping or sensitivity analyses.
Study selection will follow the predefined eligibility criteria. Titles and abstracts obtained from the database searches will be screened, and full texts of potentially relevant records will be assessed for inclusion. The review focuses exclusively on observational studies (case-control, cohort) that report original cortisol data in at-risk or diagnosed psychosis populations.
For each included study, the review will extract key study characteristics and outcome data relevant to the association between cortisol and psychosis. Extracted information will include study design, population characteristics (including risk status or diagnosis), cortisol measurement details (assay, timing, biological matrix), comparator groups, and main findings as reported by the study authors.
To assess methodological quality and potential bias, the review will use the Joanna Briggs Institute (JBI) methodology appropriate for observational designs. JBI tools will be applied to appraise domains such as selection bias, measurement reliability, confounding, and completeness of outcome reporting.
To evaluate the overall certainty of evidence across studies addressing each outcome, the review will apply the Grading of Recommendations Assessment, Development, and Evaluation (GRADE) framework. GRADE assessments will consider study limitations, consistency, directness, precision, and risk of publication bias to determine confidence in effect estimates.
The review will synthesise evidence on the association between cortisol levels and psychosis, and on cortisol levels in at-risk groups compared with relevant comparators. Where studies are sufficiently homogenous in design, population, and cortisol measurement, quantitative synthesis (meta-analysis) may be considered; otherwise, findings will be presented narratively.
Subgroup or sensitivity analyses may be explored based on factors reported in the included studies, such as timing of cortisol sampling (diurnal measures, awakening response, single time-point), biological matrix (saliva, serum, plasma), and specific risk categorizations (clinical high risk, prodromal, ultra-high risk). Any planned statistical approaches will depend on the characteristics and availability of data from included studies.
No ethical approval is required for this systematic review protocol, as the review involves synthesis of published data. The review is ongoing, and the authors plan to disseminate the findings through submission to a peer-reviewed journal.
Study registration details are provided: PROSPERO, CRD420251136994 (registered 6 October 2025).